Progress and Challenges in Lithium-Ion Battery Health State Research Based on Electrochemical Impedance Spectroscopy Modeling

Abstract

This paper provides a systematic review of the research progress and challenges in lithium-ion battery state-of-health (SOH) assessment based on electrochemical impedance spectroscopy (EIS). The study notes that SOH, as a core metric for assessing battery degradation and remaining lifespan, is evaluated through parameters such as capacity decay or changes in internal resistance. Current SOH estimation methods are primarily categorized into direct measurement methods, model-driven methods, and data-driven methods. Among these, EIS-based methods have emerged as a research hot spot due to their advantages of being fast, noninvasive, and capable of reflecting changes in internal electrochemical reactions. Model-driven methods construct equivalent circuit models (ECMs) to fit EIS data, extract parameters, and analyze aging mechanisms; data-driven methods learn features from historical data and combine algorithms to predict SOH. The article also compares the advantages and disadvantages of the two methods and proposes a hybrid method combining mechanisms and data as a future development direction. The differences in model errors among various studies are relatively small, with most falling within 2%, and several models exhibiting errors around 0.3%, indicating overall stable performance. In contrast, for neural networks, the root-mean-square error (RMSE) varies significantly across different models, ranging from 1.12% to 5.29%. This suggests substantial disparities in the fitting capabilities of different neural network architectures, resulting in relatively poor model performance stability.